The aims of this chapter are to consider human error within the context of system design and to put forward some suggestions concerning its management in workplace applications. The chapter presents error classification, causes of error, consequences of error and reduction of human error.

This chapter reviews development in the following fields: situation awareness; skill based performance; rule based performance; knowledge based performance; naturalistic decision making; skilled long term working memory; prospective memory; implicit memory and cognitive aspects of performance.

Automation will have an increasing influence on technological systems where people are responsible for vigilance functions. However, both currently and in the future, personnel will retain their critical role in monitoring, interpretation and decision-making. One needs to ensure that personnel will be able to operate in a manner and environment that will allow them to be capable of high levels of vigilance and that this performance can be sustained during working periods. Vigilance enabling conditions need to be addressed from the initial operational requirements, the specification of job and work arrangements, and the design of the console and the physical operating environment. This needs to be followed up with what kind of people and skills are appropriate, and how they will be selected, trained and managed on an ongoing basis. An important part of the technology system implementation is how to ensure that technology facilitates vigilance rather than merely assuming some of the functions previously done by people and reducing it. It may well be that the role of people within a particular system is redefined and their responsibilities changed. However, this needs to be done in a manner that results in the creation of job roles and work processes that facilitate vigilance and operator decision-making. The cost of lowered vigilance levels and the consequences of problems in overall system performance can otherwise be extremely high.

In this chapter, situation awareness (SA) is discussed. SA of an operator can be affected by the design of systems interactions within the context of use is an important safety issue. People in control of complex systems interact and operate using a remarkable cognitive process which requires the creation and maintenance of SA. An SA process model was introduced to model the process of a human operator acquiring and maintaining SA in situ. A generic SA interaction model of a typical control system and this represents both the human and technical factors that can affect the SA of people in control.

Teams, both military and civil, operate in complex and dynamic control room environments. Such teams include those working in military headquarters, nuclear power plant control rooms, on the bridges of large ships and in train operation control centres. Although the functions of the teams and the roles of the individuals within them may vary, the quality of their teamwork behaviour and performance is often influenced by a set of common factors. For many of these teams, the day-to-day operation of the control room involves overseeing systems monitoring the environment, the conduct of safety checks and the testing of equipment. In these circumstances, decision-making and team co-ordination whilst important, is relatively proceduralised.

Control rooms provide places for staff to undertake supervision and control of complex systems. Often, staff are removed from the actual environment and must monitor the system through symbolic displays, rather than direct observation. To acquire the appropriate skills an operator or system supervisor must learn how to co-ordinate actions that can be taken in the control room to deal with different circumstances that arise. Control room design decisions prescribe how the world is represented to operating staff and, therefore, substantially influence what has to be trained.

The major classical reviews of the human operator in process control are now over 25 years old (Edwards and Lees, 1973; 1974), and summarise much of what was then already known about the manual control of large complex industrial systems. They contain many hundreds of references to the role of the human operator, but the advance of automation means that the relation of humans and machines in industrial tasks needs to be re-examined.

Task analysis is widely used (and sometimes misused) to examine task performance and to assess particular features of a work situation that may have an impact on performance, such as interfaces, job aids, procedures, work organisation or training. The tasks themselves can vary considerably from short, simple routine tasks to much more complex team-based tasks that involve high-level cognitive reasoning and decision-making. Therefore, from the outset of any consideration of task analysis methods, it is important to appreciate the implications of attempting to examine several different potential issues over such a wide range of tasks. In particular, it is necessary to accept that it is unrealistic to expect that a specific method of investigation and analysis will suit all potential task analysis situations. Accordingly, the task analysis process should be seen as the selection and utilisation of a sub-set of potential methods that will be most appropriate for a particular application. Further, there are a number of practical issues associated with carrying out a task analysis and these will also be considered in this chapter.

This chapter discusses two approaches for enhancing control room teamwork. The first part of the chapter discusses a number of contemporary techniques which have been utilised successfully for team training. The second part of the chapter highlights design imperatives for building effective team support technology. The rapid advances made in team training over the past 15 years have resulted primarily from the explicit recognition of the contribution that teamworking can make to the effectiveness with which complex tasks are undertaken. Traditional approaches to team training are often founded on the assumption that if every team member does their own job well, then the 'team' as a whole will function effectively. Such approaches are consequently geared towards training a collection of individuals to do the tasks specifically allocated to their role the focus thus being on training taskwork. This approach can work well where individuals' tasks are tightly defined, stable, and self contained, e.g. for workers on a manufacturing production line. Perhaps unsurprisingly, this approach also renders team members poorly equipped to deal with ill-defined, complex, shifting tasks, which require significant co-ordination with other team members in order for the tasks to be achieved effectively. Such conditions tend to be the norm for teams operating within control rooms. Recent advances in team training have focused on training the members of a team together as a composite unit, with implicit and explicit interactions between team members being targeted specifically for practice and enhancement. Increasingly, such training is accounting for the complexities of the naturalistic environments in which teams must operate. This shift of emphasis requires a different way of thinking about how tasks are undertaken by a collection of individuals. In addition, it also requires that individual task demands are considered within the broader context of the interpersonal infrastructure, communication and co-ordination requirements of the team as a whole. Thus, the focus of these more recent approaches is on training teamwork.

It would seem relevant and appropriate to observe the activities of personnel in control rooms prior to designing new technologies to be placed in them, altering the configuration of equipment and layout, or changing the assignment of staff responsibilities and tasks. Numerous approaches have been outlined for organising and shaping these observations, most noticeably task analysis. These can support the analyst to delineate the tasks and functions of individuals, to examine the role of new technologies in the accomplishment of activities and to identify what information is available, where, and how this information should be communicated. Typically, these approaches focus on the formal tasks and activities, or at least, what can be made explicit. In this chapter the tacit and social practices on which control room work relies will be considered. Observations of work and interaction in one particular setting, station control rooms in London Underground, will be reported. This contributes to the growing number of naturalistic studies of control rooms and similar settings. The details of the practices revealed by the analysis could contribute to the development of technologies for the control room and more generic kinds of systems. In particular, we seek to re-examine how individuals 'monitor' the surrounding domain and the activities of their colleagues. We look at how the personnel utilise the available technologies that are a combination of audio, visual and computational systems and how, in concert with colleagues, they monitor the happenings in the world beyond the control room. Finally, the ways in which such an analysis can provide insights into how prototypes technologies could best be deployed in organisational settings will be examined. The chapter will conclude by briefly outlining some recent technological developments that aim to support the work of control room personnel.

The aim of the work discussed here was to produce a framework for evaluation of the human factors (ergonomics) issues associated with the introduction of new, integrated control centres within the UK railway network. The framework took the form of a measurement package Railway Ergonomics Control Assessment Package (RECAP) (Cordiner, 2000). The intention of RECAP was to enable the evaluation of: system performance in network control (including hardware, software, environment, systems and organisation); consequences for people involved in control operations; long term changes and benefits. This chapter describes the process of developing RECAP, highlighting the different methodological approaches incorporated, and identifying the relative suitability of these approaches. In particular, the practical issues associated with implementing such a tool in a large, geographically dispersed organisation that currently uses a number of different system generations and types are discussed. A generic model of the process of developing such a tool is presented, and the way in which this model was applied in the railway control domain discussed. The chapter concludes with some practical guidelines derived from the experience of measuring and analysing human factors in railway control.

It is these very changes that have brought the issue of 'ergonomics' - matching systems to people to the forefront. Systems that rely on human intervention will fail if the demands they make on their operators lie outside their capability. On the other hand, systems that can run automatically will also 'fall over' if operators do not understand how to intervene when control systems go haywire. The risks of overall systems failure are now too great for the human factor to be ignored systems designers need to take account of people-related issues and make sure that human factors are given as much attention as mechanical and electrical elements. Whilst this chapter concentrates on the role of the operator in the control room there are other 'users' of the system whose ergonomic requirements need to be considered. Lack of attention to matters such as ease of equipment maintenance, field equipment design and training will all undermine overall performance just as surely as 'loose wires' and 'bugs' in software.

This chapter relates to the design of control room alarm systems. Alarm systems are found on many user interfaces to large systems, e.g. in control rooms of power stations or chemical plants, in control centres for railway, road, air traffic or military systems, in aircraft cockpits, on bridges of ships, etc. Alarm systems are important because they are systems that provide stimulus typically an audible warning to make the operator aware of an operational problem. They should direct the operator's attention to an abnormal situation so that s/he can take preventative action. Modern processes and systems are often fitted with sophisticated protective systems to prevent hazard to people or major damage. However, action by operators is often very important to correct minor problems before they escalate into major disturbances and also to avoid unforeseen combinations of events where unprotected risks can arise. These major accidents provide obvious examples of alarm system failings. However, there is also strong evidence that large numbers of smaller and less obvious difficulties with alarm systems can have a very significant financial impact (Andow, 1998; Bransby and Jenkinson, 1998). This chapter provides an introduction on how to develop alarm systems to reduce these problems. It draws very heavily on an HSE Research Report on alarms (Bransby and Jenkinson, 1998) and the EEMUA Alarm Systems Guide (EEMUA, 1999). The latter document provides much more detailed and comprehensive guidance. In order to emphasise practical application, many of the ideas in the chapter are presented through industrial examples.

As process control plants become more complex, the demands on the operators increase. Operators need information about the process, such as the production goal, the history and ongoing maintenance tasks, as well as experience in operating a plant. It is common practice in most process plant control rooms to provide decision support systems for operators. These decision support systems help and support the operators in providing action decision help, state estimation and diagnosis. The chapter gives an introduction to decision support systems for use in process plants. Factors for designing useful systems for decision support, such as situations in which the operators need support and the different types of decision support, are discussed. The latter part of this chapter provides a presentation of a case-based reasoning approach for decision support that has been partly implemented with successful results. The proposed decision support system also integrates a three-dimensional (3D) visualisation to improve the operator's ability to diagnose and make decisions. Typically, new multimedia and visualisation techniques are used to improve existing decision support systems instead of designing new systems to utilise already existing knowledge such as individual experience and to present more intelligent information.

Focus here is on the analysis of train controllers' work and workload to determine the effects of introducing a more automated control system. The project was conducted in Melbourne, Australia, at 'CENTROL' the control centre for all trains in the state of Victoria, except trains on the Melbourne suburban passenger network. CENTROL traffic includes some long-distance passenger trains but the majority are freight trains. By international standards, traffic levels are light to moderate. One of the main control systems used at CENTROL to maintain safe separations of trains was a paper-based system termed 'Train Orders' (TO). This system relied entirely on transmission of information between controllers and field staff (usually train drivers), by means of radio or telephone conversations. A more automated control system termed 'Alternative Safe Working (ASW)' was in the process of being implemented in some areas. When fully implemented, the ASW system was expected to cover a large part of the state, with each controller work-station handling a larger area and more traffic than with TO. This was expected to require fewer controllers, based on the assumption that controller workload per train controlled would be much lower with the more automated ASW system than with TO. However, the extent of the presumed decrease in workload was uncertain, and controller workload was seen as a factor limiting the amount of territory which should be allocated to each work-station.

The privatisation of the utility sector has brought enormous changes in the way these markets operate and consumers buy their services. The electricity industry comprises a large number of private companies competing fiercely to sell electricity directly to industrial, commercial and domestic consumers. The generators that provide the power, the energy and the system ancillary services equally compete on price to produce electricity competitively into the market. This has brought with it new opportunities in the way electricity power plants need to be managed and controlled. Since privatisation, the delivery of power, energy and other ancillary services has been subject to a number of complex contractual obligations and monitoring procedures and these are changing continually. The pressure of competitive pricing places a greater emphasis on plant availability, generation efficiency, plant life and maintenance costs. The chapter describes the authors' experience of how their company has responded in this arena. Initially, it describes a number of plant management and control functions that were previously performed manually and have been automated. This has involved detailed analysis and re-design of the operator tasks. It has also required the design of many new operator interfaces and automatic control functions.

A human-centred approach to the design of work places in the railway industry is essential if railways are to retain their lead over other modes of transport in terms of safety, timeliness of the transport product, level of passenger care and energy efficiency. In this chapter, a number of case studies are presented that illustrate some of the principles of human-centred systems design, as applied to railways. The case studies relate to the handling and management of information in the context of complex systems composed of many critical sub-systems. They are used to describe situations where human-centred approaches could be applied. Railway undertakings and manufacturers of railway equipment pursue a variety of strategies for the presentation and handling of safety critical and other information.

This chapter has attempted to illustrate how the disciplines of human factors design are being applied to hitherto very traditional areas of naval engineering. Human factors engineering is now well established as an invaluable tool in the quest to develop faster and more efficient warships. More than ever before, achieving budget constraints is the main target for new ship designs. This extends beyond the expense of initial construction to the through-life costs. Personnel present the major through-life cost and therefore any reduction in ship's crew will provide substantial dividends in the battle to reduce expenditure. A major step towards achieving this aim is to control the whole ship with an Integrated Platform Management System (IPMS). However, on a warship this single network will require supreme survivability through extreme system flexibility. The issue survivability goes beyond the use of back-up cable routes and in the authors' views requires a total rethink of human-control interface philosophy.